See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/257031106

The initial release of zinc and aluminum from non-treated Galvalume and the
formation of corrosion products in chloride containing media

Article  in  Applied Surface Science · March 2012

DOI: 10.1016/j.apsusc.2011.12.112

CITATIONS READS

37 211

6 authors, including:

Xian Zhang Nam Thanh Vu

KTH Royal Institute of Technology Halliburton
8 PUBLICATIONS   312 CITATIONS    11 PUBLICATIONS   434 CITATIONS   

SEE PROFILE SEE PROFILE

Polina Volovitch Christofer Leygraf

Chimie ParisTech KTH Royal Institute of Technology
113 PUBLICATIONS   2,612 CITATIONS    338 PUBLICATIONS   12,292 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Petronas Project View project

Characterisation of Passive Films of Advanced Alloys View project

All content following this page was uploaded by Nam Thanh Vu on 09 May 2022.

The user has requested enhancement of the downloaded file.

 Applied Surface Science 258 (2012) 4351–4359

Contents lists available at SciVerse ScienceDirect

Applied Surface Science

journal homepage: www.elsevier.com/locate/apsusc

The initial release of zinc and aluminum from non-treated Galvalume and the
formation of corrosion products in chloride containing media
Xian Zhang a , Thanh-Nam Vu b , P. Volovitch b , C. Leygraf a , K. Ogle b , I. Odnevall Wallinder a,∗
a
Div. Surface and Corrosion Science, KTH, School of Chemical Science and Engineering, Dr. Kristinas v. 51, SE-100 44 Stockholm, Sweden
b
Laboratoire de Physico-chimie des Surfaces, ENSCP-CNRS (UMR7045), Ecole Nationale Supérieure de Chimie de Paris 11, rue Pierre et Marie Curie 75005 Paris, France

a r t i c l e i n f o a b s t r a c t

Article history: This study explores the initial release of zinc and aluminum from non-treated Galvalume and the par-
Received 6 October 2011 allel formation of corrosion products when exposed to synthetic seawater and rainwater of different
Received in revised form chloride content. Comparisons were made with long-term field exposures at non-sheltered marine
22 December 2011
conditions. Observed release rates from short-term conditions agree qualitatively with the long-term
Accepted 24 December 2011
findings with a selective release of zinc over aluminum. The release and corrosion processes were
Available online 30 December 2011
intertwined through the formation of corrosion products with properties that influence the long-term
release process. Prior to exposure, Al2 O3 dominated the entire surface, and was subject to local destruc-
Keywords:
Galvalume
tion upon interaction with chloride ions. As a consequence Al2 O3 was gradually replaced and covered
Metal release by zinc-rich corrosion products primarily in interdendritic areas during the first year of marine expo-
Corrosion products sure. This was followed by the gradual formation and integration of aluminum-rich corrosion products,
Chlorides reflected by an increased zinc release rate during the first year, followed by a gradually decreased
rate during subsequent years. The importance of Al2 O3 was also evident in deaerated synthetic rain-
water or seawater, where the formation of Al2 O3 was presumably hindered. In synthetic rain water
this resulted in a higher ratio between released aluminum and zinc compared with non-deaerated
conditions.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction selective release of the less noble element leaving a surface

Galvalume (other trade names, e.g. Zincalume, Aluzink, Aluz-
inc) is a commercial aluminum/zinc coating (55% Al–43.4% Zn–1.6%
Si, w/w/w [1]) applied on steel for galvanic protection, also at
cut-edges. From a corrosion perspective, the barrier capacity
of Galvalume on steel is enhanced compared with galvanized
steel (at equivalent thicknesses) in marine and industrial atmo-
spheres [2]. Galvalume and other zinc–aluminum alloys are widely
used in many applications such as electric motor housings,
door plates, window drive rails, electronics boxes, roofs, air-
conditioning equipment, appliances, and automotive under-body
parts or high temperature applications [3,4]. Several investigations
have described the spangled microstructure of Galvalume dom-
inated by Al-rich dendritic areas (64.5%Al–35.3%Zn–0.2%Si) and
interdendritic Zn-rich regions (20.4%Al–78.9%Zn–0.7% Si) contain-
ing Si [1,5] in which corrosion products initially form and evolve
[1,6–10]. The corrosion process results also in selective release
of alloy components [11]. These processes involve typically the
∗ Corresponding author. Tel.: +46 8 790 6621; fax: +46 8 208284.
E-mail address: ingero@kth.se (I.O. Wallinder).
enriched in the nobler element. As Zn and Al are both very active
metals, the local environment will possibly determine whether Al
or Zn is the active element. The material will also change as corro-
sion products evolve on the surface and selective release leads to
an enrichment of one or more alloy components in the underlying
alloy phase [12–16].
The objectives of this paper are to (i) study the initial release
of zinc and aluminum from non-treated Galvalume and the forma-
tion of corrosion products when exposed to synthetic rain water
of different chloride content and seawater at non-deaerated lab-
oratory conditions (two different set-ups), and their relevance for
long-term atmospheric outdoor conditions, and (ii) investigate dif-
ferences at deaerated conditions at open circuit conditions and at
an applied potential from a corrosion product and metal release
perspective, effects possibly taking place in oxygen depleted zones
such as crevices, and in scratches not able to re-passivate. Two dif-
ferent experimental approaches to monitor metal release (chemical
and electrochemical processes) have been applied: (i) in situ anal-
ysis of zinc and aluminum released from Galvalume held under
electrochemical control (OCP and −400 mV vs. NHE) in freshly flow-
ing media by means of inductively coupled plasma optical emission
spectroscopy (AESEC), (ii) ex situ analysis of released zinc and

0169-4332/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.apsusc.2011.12.112
4352 X. Zhang et al. / Applied Surface Science 258 (2012) 4351–4359
Table 1
Chemical composition (mg L−1 ) of synthetic rainwater [17] with different chloride content.
Synthetic rainwater NH4 NO3 Na2 SO4
[Cl− ] = 0.01 mM 0.64 0.712
[Cl− ] = 0.3 mM 0.64 0.712
aluminum by means of graphite furnace atomic absorption spec-
troscopy (GF-AAS) at specific time periods and solution volumes of
the test media at controlled conditions (temperature, agitation).
2. Material and methods
2.1. Material and test media
Galvalume, an Al–Zn coating (55%Al–43.4%Zn–1.6%Si by weight)
with a thickness of 25 ␮m on steel was subject for investigation.
Metal release studies were conducted in three different test
media including synthetic rain water (pH 4.4.) [17] with modi-
fied chloride content (0.01 and 0.3 mM) and synthetic seawater
(pH 8.1) [18]. The chemical composition of each media is given
in Tables 1 and 2. These acid and chloride containing media
were selected to enable comparison with atmospheric field data,
and be of relevance for automotive applications. All media were
prepared from analytical grade chemicals and ultra-pure water
(18.2 M cm). The pH of the test media was controlled prior to
sample exposure and kept at 4.35 ± 0.05 (if necessary adjusted by
the addition of 70–80 ␮L of 5% NaOH) and at 8.1 ± 0.05 (no pH
adjustment needed) for rainwater and seawater, respectively. All
laboratory vessels used for solution preparation were acid cleaned
in 10% HNO3 for 24 h, rinsed four times in ultra-pure water and
dried under ambient laboratory conditions.
2.2. Flow-cell tests
The AESEC (atomic emission spectroelectrochemistry) tech-
nique was employed for real-time measurements of released
metals as a function of time downstream from an electrochemical
flow cell at open-circuit potential (OCP) and at an applied anodic
potential (−400 mV vs. NHE). In this study a sample area of 0.5 cm2
was exposed to the continuously renewed flowing test media for
3000 s (50 min). The evolution of the emission signals of ICP-OES
and the open circuit potential were measured continuously and
recorded after averaging over a user defined integration period, in
this work set for 1 s (each measured value of intensity corresponds
to the average of 250,000 data points). The release rates of metals
downstream from the electrochemical flow cell were calculated
based on the measured metal concentrations, Cm , times the flow
rate per surface area. The detection was performed using wave-
length 213.85 nm (Zn) and 167.08 nm (Al) on the polychromator of
Ultima 2C Horiba JobinYvon ICP-OES spectrometer (focal distance
50 cm). The electrolyte flow rate was at 3.0 mL min−1 . The detection
limits in tested conditions were 5.1 ␮g L−1 for zinc and 4.2 ␮g L−1
for aluminum respectively. The signals of other 26 elements were
recorded by polychromator in order to better correct the possible
background variation with flow or gas formation. Applied potential
measurements were conducted using a potentiostat (EG&G Prince-
ton Applied Electronics M273A) functioning in potentiostatic mode
combined with ICP-OES.
Table 2
Chemical composition (mg L−1 ) of synthetic seawater ASTM D1141-98 [18].
Synthetic seawater NaCl MgCl2 Na2 SO4
[Cl− ] = 560 mM 24.5 5.20 4.09
K2 SO4 MgSO4 ·7H2 O CaCl2 ·2H2 O 95% H2 SO4 NaCl
0.270 1.229 0.735 11.7 ␮L
0.270 1.229 0.735 11.7 ␮L 16.948
Prior to the experiment, for all deaerated conditions, solu-
tions were deaerated by nitrogen flow at ambient temperature for
10 min. Prior to exposure, the samples were cleaned with ethanol
and ultra pure water (MilliQ, 18.2 M cm).
2.3. Immersion tests
Samples for immersion tests were cut to a dimension of
4 cm × 0.5 cm. The reverse side and the edges of the samples were
sealed with a transparent non-metal containing lacquer three times
to receive a defined exposed geometric surface area of 2 cm2 . This
lacquer has previously been shown to act as an efficient barrier
under similar immersion conditions [19]. Triplicate samples were
exposed in the different test media for the different exposure peri-
ods. Prior to exposure, the samples were cleaned with ethanol and
ultra-pure water (MilliQ, 18.2 M´ cm).
Each sample was mounted with double sided tape in
a centered position at the bottom of polypropylene boxes
(29 cm × 6 cm × 5 cm) used as exposure vessels. Four boxes in total,
triplicate samples and one blank (no sample), were exposed in par-
allel using a bi-linear shaking table providing a gentle agitation that
moved the solution over the sample surface (set at an angle of 12◦
and 30 cycles/min in order to simulate the flow-cell test conditions
to some extent). Similar exposure time periods as investigated in
the flow cell test were selected for the immersion experiments.
Three short time periods were therefore selected, 600 s (10 min),
1800 s (30 min) and 3000 s (50 min). The specific solution volumes
were based on the flow rate of solution electrolyte, the exposure
time and the surface area of the sample used in the flow cell experi-
ments. For example, in the flow cell test, 10 min (600 s) of exposure
would correspond to 30 mL of solution passing over the sample
area of 0.5 cm2 . Translated to the immersion experiments, this
would correspond to specific solution volumes of 120 mL (10 min),
360 mL (30 min) and 600 mL (50 min). Prior to exposure, the spe-
cific solution volume was transferred into the box and covered with
parafilm.
After exposure, the solution was transferred to a storage flask
and acidified to a pH less than 2 using suprapur 65% HNO3 (120 ␮L-
rainwater; 150 ␮L-seawater). This is a standard procedure prior
to total metal analysis and it serves to preserve the solution and
to completely dissociate metal complexes and avoid any metal
adsorption onto the inner walls of the storage vessel. The exposed
samples were rinsed in ultrapure water and dried in ambient air
before packed in KleenexTM tissue wipes and stored for surface
characterization.
Total concentrations of released zinc and aluminum in rainwa-
ter samples were analyzed by means of graphite furnace atomic
absorption spectroscopy (GF-AAS) at standard operational condi-
tions. Three replicate readings were made for each sample and
control samples were run consecutively during the analysis (every
10th sample). The limits of detection (defined as three times the
standard deviation between replicate readings of blank samples)
were 2.5/2 ␮g L−1 for aluminum and 1/1.5 ␮g L−1 for zinc in the rain
CaCl2 KCl NaHCO3 KBr H3 BO3 SrCl2 NaF
1.16 0.695 0.201 0.101 0.027 0.025 0.003
 X. Zhang et al. / Applied Surface Science 258 (2012) 4351–4359 4353

Fig. 1. Release rates of zinc and aluminum from Galvalume exposed in synthetic
rainwater (0.01 mM, 0.3 mM Cl− ) and synthetic sea water (560 mM Cl− ) during
immersion testing. Each marker corresponds to the mean value of three replicate
samples. (Solid lines are only used as guides.)

waters of chloride concentrations 0.01/0.3 mM, respectively. The

limit of detection for aluminum in artificial sea water was 1 ␮g L−1 .
Analysis of released zinc in seawater samples was conducted using
the flame mode with a limit of detection of 1.5 ␮g L−1 . Selected
samples were validated by analyses also by means of GF-AAS.

2.4. Long-term field exposure

Laboratory findings were compared with long-term field data

generated for non-treated Galvalume exposed at unsheltered con-
ditions at 45◦ from the horizontal facing south in the marine site
of Brest, France for 5 years (SO2 < 3 ␮g m−3 , 450 mm y−1 ), in agree-
ment to the ISO 9226 standardized exposure condition for corrosion
rate and runoff rate measurements [20]. All samples were exposed
5–10 m from the waterline [21]. Detailed information of the test
site is given elsewhere [21,22]. Fig. 2. Spontaneous release rates of zinc and aluminum for Galvalume exposed
to non-deaerated (a) synthetic rain water (0.01 mM Cl− ), (b) synthetic rain water
(0.3 mM Cl− ), and (c) synthetic seawater (560 mM Cl− ).
2.5. Corrosion product formation

A multitude of highly surface sensitive and nondestructive
analytical methods were adopted for detailed analysis of corro-
sion product formation. The approach includes scanning electron
microscopy (SEM) for morphological investigations, infrared (IRAS)
and confocal Raman microspectroscopy (CRM) for information of
functional surface groups and possibly their lateral distribution,
surface sensitive grazing incidence X-ray diffraction (GIXRD) for
detection of crystalline phases, and X-ray photoelectron spec-
troscopy (XPS) and Auger electron spectroscopy (AES) for elemental
compositional and chemical state information within the first few
nanometers of the surface.
2.5.1. Scanning electron microscopy/Energy dispersive X-ray
analysis (SEM/EDS)
Galvalume surface morphology and composition analysis were
obtained using a table-top SEM (Hitachi TM-1000) with an accel-
erating voltage of 15 kV, equipped with Hitachi EDS facility.
2.5.2. Infrared reflection absorption spectroscopy (IRAS)
The IRAS spectra were carried out by using a commercial Digilab
4.0 Pro FTIR spectrometer with 1024 scans at a resolution of 4 cm−1
in absorbance units (−log(R/R0 )), where R is the reflectance of the
exposed sample and R0 the reflectance of unexposed sample [23].
2.5.3. Grazing incident X-ray diffraction (GIXRD)
GIXRD was performed in X’pert PRO PANALYTICAL system,
equipped with an X-ray mirror (Cu K␣ radiation) and a 0.27◦ paral-
lel plate collimator on the diffracted side. Scanning was generated
on 1 cm × 1 cm surface area at a grazing angle of 88◦ vs. the surface.
2.5.4. Confocal Raman microspectroscopy (CRM)
The confocal Raman measurements were carried out with a
WITec alpha 300 system, equipped with a laser source of wave-
length 532 nm. The integration time per Raman spectrum was in
the order of 50 ms. Measurements were obtained with a Nikon
objective, Nikon NA0.9 NGC, together with a pinhole with 100 ␮m
diameter. The Raman spectra were produced in the scanning area
with a lateral resolution around 300 nm and a vertical resolution
around 2 ␮m.
2.5.5. X-ray photoelectron spectroscopy/Auger electron
spectroscopy (XPS/AES)
Analysis of chemical composition of the outermost surface layer
was performed with a XPS (Kratos AXIS HS) system combined with
an Auger unit. Wide scans and detailed scans (pass energy 20 eV) of
Al 2p, Zn 2p, Zn LMM, Cl 2p, S 2p, O 1s and C 1s were accomplished
using a monochromatic Al K␣ X-ray source (1486.6 eV) operated
at 300 W (15 kV/20 mA). Measurements were conducted at two
4354 X. Zhang et al. / Applied Surface Science 258 (2012) 4351–4359
Fig. 3. Average release rates of zinc and aluminum (top-left) and corresponding annual release Al/(Al + Zn) ratio for each year of exposure (top-right) for Galvalume exposed
to non-sheltered conditions at marine site of Brest, France for five years, and average release rates of zinc during 10 years of urban non-sheltered conditions (bottom)
(1 g y−1 m−2 equals 3.2 × 10−6 ␮g s−1 cm−2 ).
different areas of analysis, each approximately sized 0.4 mm2 . AES
mapping were performed using an acceleration voltage of 10 kV
and a beam current of 300 nA.
3. Results and discussion
3.1. Initial metal release in chloride containing media at
non-deaerated conditions
Results from the immersion tests are presented in Fig. 1. All
media revealed significantly higher release rates of zinc compared
to aluminum and increasing release rates of zinc with time. No
significant effect of chloride content on the release rate of either
zinc or aluminum was evident for the synthetic rain media of con-
stant pH. In seawater of significantly higher chloride content but
more alkaline pH compared with the rainwater media, small differ-
ences were evident after the two first time periods but significant
(factor of two) after 3000 s of exposure. These results imply that
chloride ions in seawater after a certain time period locally destroy
the protective ability of aluminum oxide at a faster rate compared
to rainwater exposure. The released concentrations of aluminum
were however for most cases very low, below or close to the limit
of detection (the marker corresponds to the limit of detection value,
no error bars displayed). Measurable amounts of aluminum were
only observed after the two longer time periods in seawater and
after the short time period in rain water (0.01 mM Cl− ). Similar
to findings in the slightly acidic (pH 4.4) rainwater of low chlo-
ride content, very low amounts of aluminum (<5.5%) were released
compared to zinc even after 3000 s in seawater (pH 8.1). Preferential
release of zinc was hence evident in both synthetic rainwater and
seawater and not pH-dependent at these conditions. At higher pH
(>10), the amphoteric behavior of Al oxide is known to be the origin
of the de-passivation of Al rich phases and causes the strong pref-
erential Al dissolution from Al-containing alloys and intermetallics
as previously reported [14,16]. Similarly, more negative corrosion
potential of de-passivated Al can be the origin of the selective Al
dissolution in Zn–Al alloys.
Real-time measurements of released metals using the AESEC
flow cell test at non-deaerated OCP conditions are presented in
Fig. 2. The OCP is also given as a function of time. Similar to the
immersion test findings, very low release rates of aluminum and
a dominance of released zinc were observed, independent of the
test media. Despite very different chloride concentrations in the
three test media, the total release currents were very similar. One
explanation may be that the corrosion rate is determined by the
cathodic reaction. However, the distribution of the total current
between the alloy components will be determined by the material.
Total release rates determined using the two different exper-
imental set-ups are compiled in Table 3 together with the
corresponding relation between released aluminum and zinc
(Al/(Al + Zn)) for the different exposure conditions. The results
clearly show a dominance of zinc release compared to aluminum
for both experimental set-ups. Quantitatively, the immersion test
resulted in significantly lower total release rates compared with
the flow cell test, in particular pronounced for seawater and the
high chloride containing rainwater media. The reason is believed
to be associated to the different exposure conditions with con-
stantly fresh media flowing over the surface in the flow cell test
 X. Zhang et al. / Applied Surface Science 258 (2012) 4351–4359 4355

Fig. 4. SEM images of Galvalume exposed for 3000 s to artificial rain water 0.01 mM, 0.3 mM Cl− and artificial sea water (560 mM Cl− ) via immersion experiments (left) and
flow-cell tests (right). The average mass ratio Al/(Al + Zn) refers to XPS compositional measurements of two separate areas (each sized 0.4 mm2 ).

and hence the continuous removal of released metals from the
media in contact with the surface, in contrast with the immersion
test with no renewal of the media within the exposure period. No
large differences in total release rates were observed between the
immersion tests despite large differences in chloride content of the
test media. Observed release rates during the flow cell tests were
approximately twice as high in rainwater of high chloride content
and in seawater compared with rainwater of low chloride content.
The measured OCP values during the flow cell test revealed a more
anodic potential in both rain water media (−650 mV vs. NHE) sug-
gesting partial passivation. During the seawater exposure the OCP
was more cathodic (−820 mV vs. NHE). These results clearly illus-
trate that metal release rates cannot be predicted from open circuit
potential measurements.
The release rates obtained under non-deaerated conditions
were in concordance with field data where the release of zinc
and aluminum were continuously monitored for Galvalume sur-
faces exposed at non-sheltered conditions in a marine environment
during five years [24]. Significantly more zinc was released com-
pared with aluminum throughout the exposure period with
an annual Al/(Al + Zn) release ratio typically varying between
0.04 ± 0.01 for individual years of exposure, Fig. 3. Similar to lab-
oratory findings, field data at the marine site (deposition rates
7–8757 mg Cl− m−2 day−1 ) showed initially increasing release rates
of zinc during the first year of exposure as a consequence of
rapid formation of zinc-rich corrosion products in primarily zinc-
rich interdendritic areas. This was followed by decreasing rates
due to the gradual formation and integration of aluminum rich

Table 3
Total release of zinc and aluminum for Galvalume after exposure (3000 s) at non-deaerated conditions in synthetic rainwater (pH 4.4) of different chloride content and in
synthetic sea water (pH 8.1,) and corresponding Al/(Al + Zn) release ratios. (LOD-limit of detection.)

Electrolyte Total Al + Zn Al/(Al + Zn) release ratio

Immersion tests (␮g cm−2 ) Flow-cell tests (␮g cm−2 ) Immersion tests Flow-cell tests

Synthetic rainwater (0.01 mM Cl− ) 4.00 11.9 Below LOD <0.01

Synthetic rainwater (0.3 mM Cl− ) 4.01 21.5 0.019 Below LOD
Synthetic seawater (560 mM Cl− ) 6.64 25.0 0.053 Below LOD
4356 X. Zhang et al. / Applied Surface Science 258 (2012) 4351–4359
Fig. 5. Average mass ratio Al/(Al + Zn) based on XPS compositional measurements
of two separate areas (each 0.4 mm2 ) after 3000 s exposure in artificial rainwater
and seawater.
corrosion products also in interdendritic areas [24]. Exposures at
urban field conditions [25] with low deposition rates of chlorides
(2 mg Cl− m−2 day−1 ), did not show any initially increased release
rates of zinc with time but rather gradually reduced rates with time
during a ten-year exposure period, Fig. 3. This may be related to
undisturbed aluminum oxides gradually hindering the release of
zinc from zinc rich corrosion products formed in the interdendritic
areas. The formation of corrosion products at laboratory conditions
is discussed in the next section.
3.2. Initial formation of corrosion products in chloride containing
media at non-deaerated conditions
Relatively similar surface morphology was observed by means
of SEM for Galvalume surfaces exposed under OCP conditions in
both flow cell and in immersion conditions, Fig. 4. In agreement
with literature findings [6–8], the SEM investigation demonstrates
a preferential formation of corrosion products in the zinc-rich inter-
dendritic areas, most pronounced after the seawater exposure.
No significant morphological differences were observed between
samples exposed to rainwater media of different chloride content
showing only slight and local corrosion. None of the exposure con-
ditions resulted in the identification of any crystalline corrosion
products by means of GIXRD, except for Al2 O3 , present in all cases.
According to literature findings, crystalline Al2 O3 is favoured by the
presence of chlorides [26,27]. The lack of crystalline zinc-containing
corrosion products may be related to the presence of amorphous
phases only, or to very small amounts and/or presence of crystalline
corrosion products within interdendritic areas impossible to detect
with GIXRD.
Analysis by means of XPS clearly demonstrated the dominance
of a thin layer of Al2 O3 in the outermost surface layer of the non-
exposed bare surface of Galvalume, Fig. 5. The presence of Al2 O3
was confirmed by means of XPS (Al 2p3/2 : 4.2 ± 0.2 eV) and sup-
ported by the occurrence of an IRAS peak at 958 cm−1 , attributed
to its amorphous form. The Al/(Al + Zn) surface mass fraction var-
ied between 0.85 and 0.99 over the surface, a ratio very different
from the bulk composition (0.55) and reported mass ratios of 0.7
and 0.2 for aluminum-rich dendritic areas and zinc-rich interden-
dritic areas, respectively [24]. The exposure in artificial rainwater
of low chloride content (0.01 mM) did not result in any signifi-
cant changes of this fraction with an average Al/(Al + Zn) mass ratio
of 0.96 (0.92–0.99) for both experimental conditions. Only small
changes were observed at the high chloride concentration (0.3 mM)
of rainwater with no differences between two experimental set-
ups (mean 0.84 (0.82–0.86)). Similar to the non-exposed surface,
both metallic (Al 2p3/2 : 72.2 ± 0.2 eV) and oxidized aluminum
peaks (74.5 ± 0.2 eV) were observed, suggesting the presence of
Al2 O3 with an estimated [28] thickness of 4–6 nm. Calculations
of the Auger parameters [29] suggested the presence of ZnO [30]
(2011.10.3 eV) and possibly Zn(OH)2 (2008.3 ± 0.3 eV) [24,30] on
surfaces exposed to the synthetic rain waters. The exposure to
seawater resulted in significant changes of the surface distribu-
tion of aluminum-rich and zinc-rich corrosion products. The mean
Al/(Al + Zn) surface mass fraction was for these surfaces reduced to
0.63 (0.58–0.68) in the case of the immersion study, and to 0.23
in the case of the flow-rate study. The metallic aluminum signal
was still observed for samples exposed in the immersion study,
but not present for the flow-cell exposure indicative the growth of
the aluminum oxide film in the later conditions. Calculated Auger
parameters suggested the presence of ZnO, Zn(OH)2 and zinc in
another oxidized phase not possible to assign (2009.5 ± 0.5 eV). In
both cases, even though more pronounced in the flow rate study,
the seawater exposure promoted the formation of zinc-rich corro-
sion products in the interdendritic areas as evident from the SEM
investigation, Fig. 4.
These findings are consistent with the metal release data show-
ing increasing release rates of zinc with time, and a faster release
rate in seawater after the longest immersion period investigated
(3000 s) for both experimental conditions. This can presumably
be explained by a faster interaction and local destruction of the
aluminum oxide (formed at high temperatures during alloy pro-
duction), present in both zinc-rich and on aluminum-rich areas,
by the high chloride content in seawater compared to the artificial
rainwaters of significantly lower chloride concentration. A larger
XPS surface fraction of zinc-rich corrosion products was however
expected, Fig. 4, for the flow cell test in rainwater of higher chlo-
ride concentration since the calculated total release rate in this case
was as high as for the seawater exposure, Table 3. This cannot be
explained but may be related to a non-uniform formation of zinc-
rich corrosion products in interdendritic areas. No chlorides were
identified by means of XPS on the surfaces exposed to any of the
rainwater media, whereas both sulfate and chloride were observed
on Galvalume exposed to seawater conditions. Previous findings
under atmospheric conditions have shown these species to be cor-
related to the local formation of basic zinc sulfates and or chlorides
in zinc-rich interdendritic areas [1,6–10].
The formation and presence of Al2 O3 suggested by XPS find-
ings were supported by IRAS observations with a band centered
at approximately 775 cm−1 , assigned as crystalline Al2 O3 [31],
and consistent with GIXRD findings. The band at 555 cm−1 sug-
gested the presence of ZnO [32], already indicated via XPS findings.
This phase was also supported by Raman measurements show-
ing a band around 570 cm−1 and a broad peak at approximately
400 cm−1 , indicative of amorphous ZnO [33]. IRAS spectra gener-
ated for surfaces exposed to seawater showed in addition a strong
band centered at approximately 3445 cm−1 , assigned to OH− , and a
band centered at 1124 cm−1 that may be assigned to SO4 2− and/or
OH− . These observations were supported by XPS confirming the
presence of sulfate, and sulfur by means of EDS. These findings
combined with the unidentified oxidized zinc compound suggested
by the Auger parameter calculation of XPS findings, suggest the
presence of an amorphous basic zinc sulfate. As also chlorides
were identified, amorphous basic zinc chlorides and/or sulfates
 X. Zhang et al. / Applied Surface Science 258 (2012) 4351–4359 4357

Fig. 6. Release rates (left) and SEM images (right) of Galvalume exposed for 3000 s to artificial rain water 0.01 mM, 0.3 mM Cl− and artificial sea water (560 mM Cl− ) at deaerated
flow-cell at an applied potential of −400 mV vs. NHE (right). The average mass ratio Al/(Al + Zn) inserted in the SEM images refers to XPS compositional measurements of
two separate areas (each 0.4 mm2 ).

cannot be excluded. IRAS bands were furthermore identified at
approximately 1599 and 1467 cm−1 , possibly assigned to CO3 2−
[6]. However, no peaks correlated to basic zinc carbonates (strong
peaks at 1060–1070 cm−1 ) [34] were identified by means of Raman
measurements.
3.3. Initial metal release and corrosion product formation in
chloride containing media at deaerated OCP and at an applied
anodic potential
In rainwater, the measured OCP values under anaerobic con-
ditions were more cathodic (−850 mV vs. NHE) as compared to
non-deaerated conditions (−650 mV vs. NHE) in the same media.
Curiously, no differences were observed between non-deaerated
and deaerated conditions in seawater (−850 mV vs. NHE). Con-
tradictory to findings under non-deaerated conditions, the higher
concentrations of chlorides in the rainwater and in seawater did not
result in a higher release rate of metals as compared with rainwater
of low chloride content. The data is summarized in Table 4. Simi-
lar to non-deaerated conditions, zinc was predominantly released
compared to aluminum in all media although with a higher pro-
portion of released aluminum compared to zinc.
It is of interest to see how the distribution of dissolution between
elements varies at higher corrosion rates. This is of significance in
confined zones were the total dissolution rate of the coating may
be driven by galvanic couples with neighboring surfaces or in the
presence of a scratch or whole to the steel substrate. An applied
anodic potential (−400 mV vs. NHE) was applied and resulted as
expected in significantly higher total metal release rates, as shown
in Table 4 and Fig. 6 (left). Note that the solutions were deaerated
such that all anodic dissolution current is driven by the poten-
tiostat. Zinc was predominantly released in the case of rainwater
exposure with non-significant released amounts of aluminum. In
seawater, the situation was the opposite with aluminum released
to a very large extent, almost 56% of the total amount of zinc and
aluminum. A heavily corroded surface with significantly corroded
aluminum-rich dendritic areas was observed by means of SEM at
these conditions, Fig. 6 (right), effects not seen at OCP conditions,
cf. Fig. 4, can be probably related with the high increase of the sur-
face pH under applied anodic potential due to the very high Zn
dissolution rate (1000 times higher than at OCP).
According to XPS, the mass distribution of aluminum- and
zinc-rich corrosion products (Al/Al + Zn) on the surface changed
from 0.75(0.57–0.93) to 0.82(0.77–0.87) to 0.56(0.47–0.65) for
Galvalume exposed to deaerated OCP conditions in rainwater of
low and high chloride content and seawater, respectively. Simi-
lar phases, Al2 O3 and ZnO were suggested by XPS to be present
for both exposure conditions. After exposure to sea-water, an
4358 X. Zhang et al. / Applied Surface Science 258 (2012) 4351–4359
Table 4
Total release of zinc and aluminum (conversion from jZn + jAl ) for Galvalume after flow cell exposure (3000 s) at non-deaerated and deaerated OCP conditions, and at deaerated
conditions and an applied potential (−400 mV vs. NHE) in synthetic rainwater (pH 4.4) of different chloride content and in synthetic sea water (pH 8.1), and corresponding
Al/(Al + Zn) release ratios. (LOD-limit of detection.)
Test media Total release (Al + Zn)
Flow-cell tests Flow-cell tests
(non-deaeration (deaeration OCP)
OCP) (␮g cm−2 ) (␮g cm−2 )
Synthetic rainwater (0.01 mM Cl− ) 11.9 9.30
Synthetic rainwater (0.3 mM Cl− ) 21.5 8.42
Synthetic seawater (560 mM Cl− ) 25.0 4.04
Fig. 7. Locally occurring corrosion products rich in S and Cl (white areas) from
AES mapping (100 ␮m × 100 ␮m) of Galvalume after 3000 s exposure at deaerated
activated (−400 mV vs. NHE) conditions in synthetic seawater.
additional aluminum peak was observed at higher bindings at
75.5 eV (Al2p3/2 ), possibly assigned as AlOOH or Al(OH)3 [35], in
agreement with findings of IRAS showing a broad band between
770 and 1000 cm−1 suggesting the additional presence of amor-
phous Al2 O3 , and/or AlOOH [9]. XPS and EDS identified sulfur as
sulfate, chlorine as chloride and sodium on the surface. Accord-
ing to AES mapping, these elements were predominantly present
in platelet-like phases formed locally in interdendritic areas as
evidenced from XPS and AES mapping, Fig. 7. These observa-
tions are consistent with findings for Galvalume exposed at
non-sheltered conditions at a marine site for five years where
ZnO, Zn5 (CO3 )2 (OH)6 , non-assigned corrosion products rich in
SO4 2− and (OH)Cl− (possibly NaZn4 Cl(OH)6 SO4 × 6H2 O [36]) and
Al(OH)3 ·1/2H2 O were identified [24]. Similar zinc rich corrosion
products have been identified on bare zinc sheet exposed in parallel
at the marine site [33].
The results clearly demonstrate that the mechanisms of zinc
and aluminum release may change radically depending upon
environmental conditions. These results could have important con-
sequences for confined zone corrosion since the gap electrolyte may
readily pass from aerobic to anaerobic conditions.
4. Conclusions
This multi-analytical study of non-treated Galvalume exposed
to short-term non-deaerated synthetic rain or seawater of
Al/(Al + Zn) ratio
Flow-cell tests Flow-cell tests Flow-cell tests Flow-cell tests
(deaeration −400 mV (non-deaeration OCP) (deaeration OCP) (deaeration, −400 mV
vs. NHE) (␮g cm−2 ) vs. NHE)
63.4 <0.01 0.043 ∼0
130 Below LOD 0.040 ∼0
4129 Below LOD 0.087 0.56
different chloride content and to a long-term natural marine out-
door environment permits the following conclusions to be drawn:
- All investigated synthetic rainwater media revealed significantly
higher release rates of zinc compared to aluminum and also
increasing release rates of zinc with time.
- Release rates were compared in two short-term exposure con-
ditions, stagnant immersion test and flow cell test. A selective
release of zinc was always observed, but the immersion test
resulted in significantly lower total release rates compared with
the flow cell test.
- During immersion conditions the total release of zinc and
aluminum was similar in all media, whereas under flowing con-
ditions the total release rate increased with chloride content.
- Results from non-deaerated short-term conditions agree qual-
itatively with long-term non-sheltered exposure in the marine
environment: a selective release of zinc was observed throughout
the five-year exposure period with the annual Al/(Al + Zn) release
ratio typically in the range 0.04 ± 0.01, also an initial increase in
zinc release rate during the first year followed by a decreasing
zinc release rate.
- This time-dependence is a consequence of the gradual build-up of
zinc-rich corrosion products during the first year, particularly in
interdendritic areas, followed by the formation and integration of
aluminum-rich corrosion products at exposure times longer than
one year.
- The only crystalline corrosion product detected was Al2 O3 . This
phase dominates the entire Galvalume surface before exposure,
and is subject to local destruction upon interaction with chloride
ions. As a consequence Al2 O3 is gradually replaced and covered
by zinc-rich corrosion products in the interdendritic parts, most
likely non-crystalline basic zinc chlorides and/or sulfates, during
the first year of marine exposure.
- Also in deaerated synthetic rainwater, zinc was selectively
released, although the ratio between released aluminum and zinc
was higher than in non-deaerated conditions. However, in deaer-
ated synthetic seawater and at an applied potential, aluminum
release dominated over zinc release, presumably because of the
hindrance of the formation of a protective Al2 O3 layer.
- In all, this study is another confirmation of the different mech-
anisms that govern metal corrosion and metal release, two
processes that are intertwined through the formation of corrosion
products with properties that influence the long-term release of
Galvalume.
Acknowledgements
Financial support from the Chinese Scholarship Council (CSC) is
gratefully acknowledged.
The authors are grateful for the financial support of the field
exposure provided by Nordic Galvanizers Association, Sweden;
Rheinzink, Germany; Saferoad, Norway and SSAB, Sweden.
 X. Zhang et al. / Applied Surface Science 258 (2012) 4351–4359
The research leading to these results has received funding from
the European Union’s Research Programme of the Research Fund for
Coal and Steel (RFCS) research programme under grant agreement
no. RFSR-CT-2009-00015.
Instrumental grants from Knut and Alice Wallenberg foundation
(XPS) and from Nils and Dorthi Troëdsson Foundation (combined
confocal Raman and AFM) are gratefully acknowledged.
The French Corrosion Institute, Brest, France, is highly acknowl-
edged for their invaluable help in collecting runoff water at the
marine site.
Valuable discussions from Dr. Ping Qiu and Dr. Klara Midander
are highly appreciated.
We also thank Dr. Gunilla Herting and MSc. David Lindström for
all field exposure AAS analytical efforts.
References
[1] A.R. Moreira, Z. Panossian, P.L. Camargo, M.F. Moreira, I.C.d. Silva, J.E.R. de Car-
valho, Zn/55Al coating microstructure and corrosion mechanism, Corros. Sci.
48 (2006) 564–576.
[2] E. Palma, J.M. Puente, M. Morcillo, The atmospheric corrosion mechanism of
55% Al–Zn coating on steel, Corros. Sci. 40 (1998) 61–68.
[3] W.-B. Chen, P. Chen, H.Y. Chen, J. Wu, W.-T. Tsai, Development of Al-containing
zinc-rich paints for corrosion resistance, Appl. Surf. Sci. 187 (2002) 154–164.
[4] U.S. Steel Corporation, GALVALUME® sheet steel, in: Technical Bulletin Con-
struction, 2005.
[5] J.H. Selverian, M.R. Notis, A.R. Marder, The microstructure of 55 w/o Al–Zn–Si
(Galvalume) hot dip coatings, Mater. Eng. 9 (1987) 133–140.
[6] A.K. Neufeld, I.S. Cole, Use of FTIR to study surface changes on metals in the early
stages of degradation, in: 13th International Corrosion Congress, Melbourne,
Australia, 2006, Paper 046.
[7] J.J. Friel, Atmospheric corrosion products on Al, Zn and AlZn metallic coatings,
Corrosion 42 (1986) 5.
[8] E. Palma, M. Morcillo, A laboratory study of the cathodic protection afforded by
55-percent Al–Zn coating in comparison with a galvanized coating, Plat. Surf.
Finish. 85 (1998) 106–109.
[9] I. Odnevall, Characterization of corrosion products formed on rain sheltered
AluzinkTM and aluminum in a rural and an urban atmosphere, in: Proc. 13th
Int. Corr. Congr, Australia, 1996, 1–8.
[10] I. Odnevall Wallinder, W. He, P.E. Augustsson, C. Leygraf, Characterisation of
black rust staining of unpassivated 55% Al–Zn alloy coatings. Effect of temper-
ature, pH and wet storage, Corros. Sci. 41 (1999) 2229–2249.
[11] M. Mokaddem, P. Volovitch, K. Ogle, The anodic dissolution of zinc and zinc
alloys in alkaline solution. I. Oxide formation on electrogalvanized steel, Elec-
trochim. Acta 55 (2010) 7867–7875.
[12] K. Ogle, A. Tomandl, N. Meddahi, M. Wolpers, The alkaline stability of phosphate
coatings I: ICP atomic emission spectroelectrochemistry, Corros. Sci. 46 (2004)
979–995.
[13] K. Ogle, M. Mokaddem, P. Volovitch, Atomic emission spectroelectrochemistry
applied to dealloying phenomena II. Selective dissolution of iron and chromium
during active-passive cycles of an austenitic stainless steel, Electrochim. Acta
55 (2010) 913–921.
[14] M. Mokaddem, P. Volovitch, F. Rechou, R. Oltra, K. Ogle, The anodic and cathodic
dissolution of Al and Al–Cu–Mg alloy, Electrochim. Acta 55 (2010) 3779–3786.
[15] K. Ogle, J. Baeyens, J. Swiatowska, P. Volovitch, Atomic emission spectroelectro-
chemistry applied to dealloying phenomena: I. The formation and dissolution of
residual copper films on stainless steel, Electrochim. Acta 54 (2009) 5163–5170.
View publication stats
4359
[16] K. Ogle, M. Serdechnova, M. Mokaddem, P. Volovitch, The cathodic
dissolution of Al, Al2Cu, and Al alloys, Electrochim. Acta 56 (2011)
1711–1718.
[17] G. Herting, I. Odnevall Wallinder, C. Leygraf, A comparison of release rates of Cr,
Ni and Fe from stainless steel alloys and the pure metals exposed to simulated
rain events, J. Electrochem. Soc. 152 (2005) B23–B29.
[18] American Society for Testing Materials, Standard Practice for the Preparation
of Substitute Ocean Water, D1141-98, 2003.
[19] G. Herting, I. Odnevall Wallinder, C. Leygraf, Corrosion-induced release of
chromium and iron from ferritic stainless steel grade AISI 430 in simulated
food contact, J. Food Eng. 87 (2008) 291–300.
[20] ISO, ISO standard 9226, Corrosion of metals and alloys – Corrosivity of atmo-
spheres – Determination of corrosion rate of standard specimens for the
evaluation of corrosivity, in, 1992, and ISO/DIS 17752: Corrosion of metals
and alloys – Procedures to determine and estimate runoff rates of metals from
materials as a result of atmospheric corrosion, 2011.
[21] J. Sandberg, I. Odnevall Wallinder, C. Leygraf, N. Le Bozec, Corrosion-induced
zinc runoff from construction materials in a marine environment, J. Elec-
trochem. Soc. 154 (2007) C120–C131.
[22] D. Lindström, I. Odnevall Wallinder, Long-term use of galvanized steel in exter-
nal applications. Aspects of patina formation, zinc runoff, barrier properties
of surface treatments, and coatings and environmental fate, Environ. Monit.
Assess. 173 (2011) 139–153.
[23] T. Aastrup, C. Leygraf, Simultaneous infrared reflection absorption spec-
troscopy and quartz crystal microbalance measurements for in situ
studies of the metal/atmosphere interface, J. Electrochem. Soc. 144 (1997)
2986–2990.
[24] P. Qiu, C. Leygraf, I.Odnevall Wallinder, Evolution of corrosion products and
metal release from Galvalume coating during short and long-term atmospheric
exposures, Mater. Chem. Phys., in press.
[25] S. Bertling, I. Odnevall Wallinder, C. Leygraf, D. Berggren Kleja, Occurrence and
fate of corrosion-induced zinc in runoff water from external structures, Sci.
Total Environ. 367 (2006) 908–923.
[26] C.L. McBee, J. Kruger, Nature of passive films on iron-chromium alloys, Elec-
trochim. Acta 17 (1972) 1337–1341.
[27] T.P. Hoar, On corrosion-resistant materials, J. Electrochem. Soc. 117 (1970)
17C–22C.
[28] B.R. Strohmeier, An ESCA method for determining the oxide thickness on alu-
minum alloys, Surf. Interface Anal. 15 (1990) 51–56.
[29] C.D. Wagner, The Auger parameter, its utility and advantages: a review, J. Elec-
tron. Spectr. Relat. Phenom. 47 (1988) 283–313.
[30] A.K. Chandra, R. Mukhopadhyay, J. Konar, T.B. Ghosh, A.K. Bhowmick, X-ray
photoelectron spectroscopy and Auger electron spectroscopy of the influence
of cations and anions of organimetallic adhesion promoters on the inter-
face between steel cord and rubber skim compounds, J. Mater. Sci. 31 (1996)
2667–2676.
[31] P. Brüesch, R. Kötz, H. Neff, L. Pietronero, Vibrational properties of Al2 O3 films
on gold, aluminum, and silicon, Phys. Rev. B 29 (1984) 4691.
[32] F. Zhu, X. Zhang, D. Persson, D. Thierry, In situ infrared reflection absorption
spectroscopy studies of confined zinc surfaces exposed under periodic wet–dry
conditions, Electrochem. Solid State Lett. 4 (2001) B19–B22.
[33] J. Hedberg, N. Le Bozec, I. Odnevall Wallinder, Spatial distribution and formation
of corrosion products in relation to zinc release for zinc sheet and coated pre-
weathered zinc at urban and a marine atmospheric conditions, Mater. Corros.
(2011), doi:10.1002/maco.201106361.
[34] M.C. Hales, R.L. Frost, Synthesis and vibrational spectroscopic characterisation
of synthetic hydrozincite and smithsonite, Polyhedron 26 (2007).
[35] Y. Zhang, Characterization of as-received hydrophobic treated AlN powder
using XPS, J. Mater. Sci. Lett. 21 (2002) 1603–1605.
[36] I. Odnevall, C. Leygraf, Formation of NaZn4 Cl(OH)6 SO4 6H2 O in a marine atmo-
sphere, Corros. Sci. 34 (1993) 1213–1229.